Binus University Repository

Mulyadi, Martin Surya

Form Peer Review "Corporate Governance, Earnings Management and Tax Management"

Rudy, Rudy

Storyboarding for Meetro: A Smart Navigation Application System for Commuter 
Line in Jabodetabek 
 
Sevenpri Candra 
School of Business Management 
Bina Nusantara University 
Jakarta 11480, Indonesia 
e-mail: seven@binus.ac.id 
Zelda Soraya 
School of Business Management 
Bina Nusantara University 
Jakarta 11480, Indonesia 
e-mail: seven@binus.ac.id 
Rudy 
School of Information Systems 
Bina Nusantara University 
Jakarta 11480, Indonesia 
e-mail: rudy@binus.edu 
even@binus.ac.id 
 
Abstract—Transportation has been long part of people life, 
especially for the urban. The transportation need not only 
applied for the local, Jakarta as a capital city must be full of 
foreigner either for business purpose or others, and they are 
need a public transportation as well.  Commuter Line 
Jabodetabek is the most preferable public transportation to 
avoid Jakarta’s traffic. The idea is to create a mobile 
application system which could easily help Commuter Line 
Jabodetabek user to get the information about timetable 
schedule, arrange the route and transit information and find 
how to reach the destination as well as how long to get there 
and how much does it cost, and visualize the route in easy-to-
understand format.  The application system is analyzed and 
designed with mobile application development approach and 
the application development with cloud computing provided by 
Application Craft, and the result of this study is prototype 
application design with cross-platform based. (Abstract) 
Keywords—commuter line; transportation; information; 
mobile application;  (key words) 
I.  INTRODUCTION 
In developed countries, the road based transportation 
systems become the government focus to develop the 
infrastructure. Jabodetabek (Jakarta, Bogor, Depok, 
Tangerang and Bekasi) is a region that consists of DKI 
Jakarta Province, the capital city of Republic of Indonesia, 
and there are 7 other regions such as Bogor City, Depok City, 
Tangerang City, and Bekasi City that surrounds Jakarta, as 
hinterland. According to [1], Figure 1 shows that there is 
increment in commuter traffic from Jakarta Metropolitan 
Area to DKI Jakarta from 2002 to 2010 about 1.5 times. But 
the high demands of commuter line Jabodetabek transport 
are not balanced with the high level of service. But the high 
demands of commuter line Jabodetabek transport are not 
balanced with the high level of service.  Proven by JICA’s 
research based on Journey experience index that shows 
Jakarta as the city with the lowest rate of commuter 
satisfaction for public transportation in major Asian cities 
(See Figure 2).  
Based on the underlying reasons, this study is devoted to 
the establishment of an idea to create such a mobile 
application (with cross-platform based and business model as 
a differentiation tool in business environment) which 
purpose is to assist Commuter Line Jabodetabek users with 
their commuter activities.  This application system could 
easily provide the information about timetable, departure 
time, and the way on how to reach the destination station, 
how long and how much does it cost to get there, including 
the transit information with the visualization of the routes in 
a simple and easy-to-understand format (using map routes or 
highlight lists).  The proposed application system is expected 
to help commuter line Jabodetabek user to get the 
information about timetable and help to plan the departure 
time to minimize the waiting time at the station, and the most 
important thing is to create the environment of using public 
transportation handled nicely, instead of making it a stressful 
experience. 
 
Figure 1. Increment of Commuter Traffic Trip Source: [2] 
 
 
Figure 2. Commuter Satisfaction of Public Transport in Major Asian Cities 
Source: [1] 
304
2017 3rd International Conference on Information Management 
978-1-5090-6306-2/17/$31.00 ©2017 IEEE
II. METHODS
The study will be started by conducting industry analysis 
which are represented by Porter’s Five Forces Model 
Analysis and Consumer Behavior Analysis. After conducting 
the analyses, the results of analyses will be used to transform 
the idea into a system design. System analysis and design 
will use mobile application development approach, which 
will be represented by Storyboarding, Server/Client Flow, 
Design UI, Online / Offline Development, Testing, signing 
certificate, and distributing the prototyped application [3], [4]. 
Mobinex, a leading provider of mobile applications and 
on-device solutions, is committed to increase revenues of 
mobile operators, content provider, media companies and 
enterprises with compelling user experiences [5]. The 
methodology was proposed in 2009 (updated in 2010) for 
developing mobile applications. The main phases of 
methodology are represented in Figure 3.
 
Figure 3. Mobinex Mobile Application Development 
Source: [5] 
III. RESULTS
Based on Mobile Application Development, the first step 
is need assessment. Needs assessment will cover customer 
segment, customer relationship, channels, value proposition, 
key activities, key resources, key partners, revenue streams, 
and cost structure. In customer segments section, Meetro will 
providing several features to support customer on each 
segment. Commuter line Jabodetabek daily user will enable 
set personal setting such as language preference, and 
bookmark the station. And business owner near user will be 
verified to minimize potential unauthorized business owner. 
In customer relationship section, Meetro will providing 
several features to engage customer relation. Automated 
service will transform into easy search station features, and 
automated campaign for business owner to promote their 
business. Co-Creation enable business owner to manage their 
own contents, and personal assistance here will transform 
into personal campaign such as welcome greeting, personal 
setting apps, and personalized promotion to commuter line 
Jabodetabek user. For Meetro’s Channel, Meetro take an 
advantage from cross-platform development to distribute 
Meetro apps in all smartphone marketplace or by simply 
scan the QR code. Meetro also will optimizing social media 
and Search Engine Optimization (SEO) to get internet citizen 
attention. 
To deliver Meetro’s value preposition, Meetro will 
available in all smartphone’s platform to give all in one 
service. Real-time information will be available in simple 
monitoring user interface. Meetro apps is free to download 
and get beneficial information, and Business owner can take 
an advantage to promote their business to commuter line 
Jabodetabek user through Meetro. Meetro Key Activities will 
focus on Problem Solving by give a solution from problem 
analysis based on consumer characteristics. In platform 
management, Meetro will keep the information stay updated, 
enable business owner to managing their content in simple 
steps. Marketing and networking activities will be focusing 
on optimizing social media and smartphone marketplace to 
get internet citizen attention. 
For Meetro’s key resources, Meetro proposed several 
systems to maintain the relationship with its resources. 
Meetro will promote human resource event such as post job 
vacancy in social media and career page; and several human 
resource activities in social media and Meetro’s page. 
Maintaining relationship with PT. KCJ using publication our 
regular meeting with PT. KCJ in our page. For Meetro’s Key 
Partners, Meetro proposed ads registration and display space 
for our business owner and ads registrant. Ads registrant and 
business owner can promote their business in profitable way. 
Ads status and billing status can be monitored by accessing 
their account in Meero’s system.
For Meetro’s Channel, Meetro take an advantage from 
advertising by providing an easy ads registration and 
payment system. Business owner verification in simple and 
quick processing also provided for business owner 
membership program. Meetro donator also would not get any 
trouble in our donating system. For Meetro’s Channel, 
Meetro take an advantage from cross-platform development 
to distribute Meetro apps in all smartphone marketplace or 
by simply scan the QR code. Meetro also will optimizing 
social media and Search Engine Optimization (SEO) to get 
internet citizen attention. 
By using mobile application development from Mobinex, 
Meetro application system will be pictured using several 
stages from storyboarding, server / clients UML flow, design 
UI, and prototype development. These steps are considered 
powerful for time efficiency on mobile application 
development and this study will be discussing storyboard 
perspective only. 
Regular Update 
App Distribution 
Signing Process 
Testing App 
Offline / Online Development 
Design 
Server / Client UML Flow 
Storyboarding 
Need Assessment 
305
1. Start Application 2. Main Screen 3. Select Depart Station
4. Select Destination Station 5. Itinerary Result
Figure 4. Meetro Intinerary Subsystem Storyboard 
306
1. Station Menu 2. Info Station 3. View Merchant
Figure 5. Info Station Storyboard 
IV. DISCUSSIONS AND CONCLUSIONS
Application system has been analyzed and designed with 
mobile application development approached in Cross-
platform based. To tear down system systems complexity, 
here storyboard will be discussed in four major subsystems 
including Itinerary subsystem, advertising subsystem, 
donation subsystems, and business owner subsystem. 
Itinerary subsystem is subsystem which handles all 
activities related to creating, and managing trip using 
commuter line Jabodetabek (see Figure 4). Flow of activities 
in itinerary subsystem is as follows: (1) Application starts the 
flash screen; (2) Meetro will show the main screen that 
provides all commuter line Jabodetabek station, there are 
several stations that not available for commuter line 
Jabodetabek (noted by red color) and several stations only 
available for commuter line for several times only (noted by 
pink color); (3) To get an itinerary, user simply selecting 
departure and destination station. On each station, there are 
also several options menu related such as timetable schedule, 
information about station, and ability to bookmark the station; 
(4) System will give an itinerary suggestion based on user 
input and some itinerary details such as fare information, 
time estimation, and total station that user needs to pass. 
As already explained, in this subsystem Meetro also 
provide additional information such as timetable schedule 
and station information on each station (see Figure 5). In 
timetable menu, user can get information related to timetable 
for each station, specifically on each line and each bound in 
simple design. Here, user can manage their trip with 
commuter line Jabodetabek comfortably. In station info 
menu, user can get information such as exit door info, toilet 
location info, and breastfeeding space facility info, disable 
access info, elevator info, also store and merchant info. In 
bookmark menu, user can bookmark the station and 
specifically on which line and which bound. User can access 
it by simply touch the bookmark button in main screen 
button. 
Business owner subsystem storyboard handles all 
activities related to creation and management of business 
owner in Meetro. Flow of activities in this subsystem is as 
follow: (1) Business owner who wish to join Meetro as the 
business owner member, simply access the business owner 
page by touch the setting button; (2) At the end of the form, 
business owner is also asked whether they want to advertise 
their business. If business owner wants to advertise their 
business, then they will be directed to Ads registration page 
to configure the advertisement and put their commitments on 
paying advertising service fee. Please note here that business 
owner required to owned business located on station or at 
least near the station. Business owner also have a privilege to 
broadcast their promotion with additional charge. Other 
business owner who doesn’t have business located near 
station but wish to promote their business within Meetro’s 
ads space, can go directly to Ads page; (3) After business 
owner submit their information, Meetro administrator will 
review it, if administrator have confidence in business owner 
information, and double check that the business located near 
station, and payment already made by business owner, then 
business owners’ merchant will be shown at station merchant 
307
info; (4) Business owner will get an account on our system, 
business owner allowed to modify their profile and 
publishing new promotion by their own. In Payment menu, 
business owner also can view their monthly payment status 
(detailed version also available). 
Advertising-subsystem storyboard handles all activities 
regarding creation and management of ads including 
payment of registered ads at the end of the period. Since 
Meetro will be built on hybrid application, her business 
owner can access Meetro both through mobile apps and 
browser, besides Meetro are consider that business owner 
might not able or interested on installing Meetro app. Here, 
Ad registrant can configure the Ad mechanism in two 
options, first the advertising fee will be based on default bid 
per click, or Ad registrant can set up their budget for an ad in 
Meetro. After ad registration is made and verification has 
been settled, system automatically campaigning the business 
on our apps.  
Ad registrant can update their ad configuration and view 
their billing status with detailed billing by simply logging in 
into their page every time they want while ad campaign is 
ongoing. At the end of the month between agreed periods, 
users need to make a payment for this ad service. For our 
merchant with ad service, they are no need to register a new 
account. They just need to make a configuration with the 
same email and Meetro will charge the ad service with 
merchant’s monthly charge.
V. CONCLUSIONS
Meetro is a feasible idea to be implemented based on 
following analysis. Meetro project is a new entry in this 
attractive industry with competitive environment, fair barrier 
to new entry with relatively low threat to the existing (unless 
new entrant can provide something useful), relatively low 
threat of substitute product, high supplier’s bargaining power, 
and high consumer’s bargaining power. In this kind of 
environment, Meetro project should be powerful enough to 
catch market attention and be more active to attract and 
maintain customer needs. 
Customer behavior analysis with performance 
satisfaction analysis results four major area, Meetro will 
focus in area with high customer expectation. The accuracy 
of information if customers need, Assurance of information 
availability related KRL Jabodetabek timetable, Ability to 
provide detailed information, and Ability to provide 
information related to KRL Jabodetabek fare cost estimation 
are the focus Meetro needs to catch customer attention on 
Market. 
Meetro Business model canvas as the tool for Meetro to 
penetrate the market. Meetro proposes several values to its 
customer such as free apps to download, available in all 
platform device, real-time commuter line Jabodetabek 
information and space for Ads to get in touch with business 
customer. In customer segments compartment, Meetro 
targeted Commuter line Jabodetabek user, Traveler, and 
Business owner near station. Relationship with customer will 
be conducted by the automated service, co-creation, and 
personal assistance. Meetro will use smartphone marketplace 
and web browser internet as its channel to deliver Meetro’s 
values. Meetro’s key activities are problem solving, platform 
management, also marketing and networking activities. Key 
resources of Meetro are human resources, PT. KCJ’s 
information, and service platform. Meetro partners including 
PT. Kereta Commuter Jabodetabek, Cross-platform tools 
(application craft and PhoneGap) and business owner near 
station. As for Meetro cost structure, main costs in Meetro is 
allocated for software and infrastructure development in the 
first year since Meetro is using several cloud services.  Last, 
to pay for all activities and costs, Meetro revenue will be 
generated from donation system, advertising, and business 
owner membership.  
Application system has been analyzed and designed with 
mobile application development approached in Cross-
platform based. Meetro system is divided into four 
subsystems: Meetro itinerary subsystem, Meetro business 
owner subsystem, Meetro advertising subsystem, and Meetro 
donation subsystem.Meetro will be implemented in three 
phases with estimated length 105 working days and each 
phase will spend 35 working days. Furthermore, there will be 
three marketing programs plan that need to be executed as 
soon as phase 1 has been fully developed and launched. 
REFERENCES
[1] Japan International Cooperation Agency, “Master Plan Study for 
Establishing Metropolitan Priority Area for Investment and Industry 
(MPA) in JABODETABEK Area,” 2012. [Online]. Available: 
https://www.jica.go.jp/english/news/press/2012/121009.html. 
[Accessed: 19-Oct-2016]. 
[2] JUTPI, “Future Public Transport Demand.,” JABODETABEK Urban 
Transportation Policy Integration (JUTPI), Jakarta - Indonesia, 2015. 
[3] Mobinex, “Mobile Application Development Methodology,” 2015. 
[Online]. Available: http://docslide.us/documents/mobile-application-
development-methodology-v3.html. [Accessed: 10-Oct-2016]. 
[4] J. W. Satzinger, R. B. Jackson, and S. D. Burd, Systems Analysis and 
Design in a Changing World, 6th Edition, 6th edition. Boston, MA: 
Course Technology, 2011. 
[5] V.-D. Pavaloaia, “Methodology Approaches Regarding Classic versus 
Mobile Enterprise Application Development,” Informatica 
Economica, vol. 17, no. 2/2013, pp. 59–72, Jun. 2013
308


English

Storyboarding for meetro: A smart navigation application system for commuter line in Jabodetabek

Anggraeni, Adilla

Form Reviewer "The Study of Consumer Behavior of Online Hotel Booking in Jakarta"

Dyah Budiastuti, Dr.

Improving information performance of schools in higher education through IT service management

Reina, Reina

 
 2017 International Conference on Information Management and Technology (ICIMTech) 
   
Organizing Committee 
 
General Chairs 
Assoc. Prof Anderes Gui Ph.D, Bina Nusantara University, Indonesia 
 
Advisory Board 
Prof. Dr. Ir. Harjanto Prabowo, MM, Bina Nusantara University, Indonesia 
Prof. Bahtiar Saleh Abbas, Bina Nusantara University, Indonesia 
Dr. Meyliana, Bina Nusantara University, Indonesia 
Dr. Ford Lumban Gaol, IEEE Indonesia Section 
 
General Committee 
Vice Chairs Dr. Yohannes Kurniawan, Bina Nusantara 
University, Indonesia 
Tanty Oktavia, Bina Nusantara University, 
Indonesia 
Secretary & Registration Cadelina Cassandra, Bina Nusantara University, 
Indonesia 
Amelia Anggraini, Bina Nusantara University, 
Indonesia 
Finance Chair & Treasurer Ratna Sari, Bina Nusantara University, Indonesia 
Program Chair Dr. Heri Pracoyo, Bina Nusantara University, 
Indonesia 
      Session Chairs Devyano Luhukay, Bina Nusantara University, 
Indonesia 
Tri Pujadi, Bina Nusantara University, Indonesia 
Evaristus Didik M, Bina Nusantara University, 
Indonesia 
sPublication Chair Fredy Jingga, Bina Nusantara University, 
Indonesia  
      Proceeding Marisa Karsen, Bina Nusantara University, 
Indonesia 
Publicity & Public Relations Chair Natalia Limantara, Bina Nusantara University, 
Indonesia 
      Web & Social Media Viriya O. H. Wongso, Bina Nusantara University, 
Indonesia 
      Graphic Designer Arilla Nurmoslim, Bina Nusantara University, 
Indonesia 
 
Proceedings of 
2017 International Conference on 
Information Management and 
Technology 
(ICIMTech) 
 
Date and Venue: 
15-17 November 2017, Melia Purosani, Special Region of Yogyakarta, 
Indonesia 
ISBN 
978-1-5386-2930-7 
IEEE Catalog Number 
CFP17H83-ART 
 
978-1-5386-2930-7/17/$31.00 ©2017 IEEE 15-17 November 2017, Melia Purosani Hotel, Yogyakarta, Indonesia 
2017 International Conference on Information Management and Technology (ICIMTech) 
Page 112 
 
 
Quadratic Classifier from Discriminant Analysis for 
Classification of Multiple Attributes Data 
(Case Study: Fertility Data Set) 
Reina Setiawan 
Computer Science Department, BINUS Graduate Program - Doctor of Computer Science  
Computer Science Department, School of Computer Science 
Bina Nusantara University 
Jakarta, Indonesia 11480 
reina@binus.edu 
 
 
Abstract—Classification is a process to group data, based on 
characteristics into related class. There are many methods in 
classification and the appropriate method is chosen based on 
nature of data. This paper focuses on classification of multiple 
attributes data using discriminant analysis. The research uses 
Fertility Data Set from UCI Machine Learning Repository with 
ten attributes of data. The experiment uses several methods of 
classification to find out the best result of performance. The 
result shows Quadratic Classifier from discriminant analysis 
has the best performance of classification around ninety-eight 
percent with the lowest errors. In summary, the appropriate 
method produces a good performance of classification and the 
quadratic classifier from discriminant analysis shows the best 
performance in multiple attributes data classification. 
Keywords—classification; discriminant analysis; quadratic 
classifier; multiple attributes  
I.  INTRODUCTION 
Classification is a process to group data, based on 
characteristics into related class. The classification process 
contains two steps; training and testing [1]. In training 
process, data is labeled based on the classes and a model is 
learned. The model is used for predicting data with unknown 
labels in testing process [2][3]. There are many methods in 
classification, such as decision tree [4], naïve bayes [5][6][7], 
k-nearest neighbor [8], support vector machine [9], and 
discriminant analysis [10].  
In recent years, many researches to compare performance 
of classification method [11][12][13][14]. The method is 
chosen based on nature of data. Nature of data with multiple 
attributes cannot be classified using linear method. This paper 
focuses on classification of multiple attributes data using 
discriminant analysis. 
The research uses Fertility Data Set from UCI Machine 
Learning Repository with ten attributes of data and a hundred 
sample of semen. The semen sample data have been analyzed 
based on criteria of the WHO 2010. Concentration of the 
sperm are related to life habits, health status, environmental 
factors, and socio-demographic data [15]. The data is 
classified using discriminant analysis method, both linear 
classifier and quadratic classifier. The reason to use linear 
classifier and quadratic classifier is to measure and to compare 
the performance of those methods. Neural network also used 
as a comparison. The experiment uses MATLAB and the 
result proves that performance of the quadratic classifier is 
better rather than performance of linear classifier and neural 
network. The result also shows the best performance of 
classification around 98% with the lowest errors. In summary, 
the appropriate method produces a good performance of 
classification and the quadratic classifier from discriminant 
analysis shows the best performance in multiple attributes data 
classification. 
II. LITERATURE REVIEW 
A. Classification 
In machine learning, classification consists of two 
processes; training process which is to learn and testing 
process which is to label the class [1]. Generally, two-thirds 
of data is part of training process and one-thirds of data is 
part of testing process. Classification model in machine 
learning can be seen in Figure 1. 
 
B. Linear Discriminant Analysis (LDA) 
Linear Discriminant Analysis (LDA) from R. A. Fisher 
uses Bayes Theorem (1). 
 
 P(i|x) = P(x|i).P(i) / ∑ P(x|j).P(j)         (1) 
LDA is based on concept of searching for a linear 
combination of variables (predictors) that best separates two 
or more classes (targets). R. A. Fisher implemented the LDA 
through Iris flower data set or Fisher iris data set. It is a 
multivariate data set consist of three classes; Iris setosa, Iris 
virginica, and Iris versicolor. Number of the data is fifty data 
per each class with four features; sepal length, sepal width, 
petal length, petal width in centimeters. Figure 2 shows 
distribution of the iris classes [16]. In the distribution, there is 
overlapping between iris virginica and iris versicolor. 
 
 
978-1-5386-2930-7/17/$31.00 ©2017 IEEE 15-17 November 2017, Melia Purosani Hotel, Yogyakarta, Indonesia 
2017 International Conference on Information Management and Technology (ICIMTech) 
Page 113 
 
 
Fig. 1. Classification model in machine learning 
 
Fig. 2. Distribution of the iris classes [16] 
It is tough to implement linear classification method to data 
with various of features or various of attributes or various of 
dimensions. Therefore, discriminant analysis is the 
appropriate method of classification. In MATLAB, classifiers 
in discriminant analysis consists of linear as a default, 
quadratic, diagLinear, diagQuadratic, pseudoLinear, and 
pseudoQuadratic. The estimation of one covariance matrix 
that cover all classes using Linear, meanwhile estimation of 
one covariance matrix in every class using Quadratic. 
DiagLinear uses the diagonal of the ‘linear’ covariance matrix 
and uses its pseudoinverse if necessary, DiagQuadratic uses 
the diagonals of the ‘quadratic’ covariance matrices and uses 
its pseudoinverses if necessary, PseudoLinear uses the 
pseudoinverse of the ‘linear’ covariance matrix if necessary, 
and PseudoQuadratic uses the pseudoinverses of the 
‘quadratic’ covariance matrices if necessary [17]. 
C. Neural Network (NN) 
Neural Network (NN) is a part of artificial intelligent that 
is adopted from the way of human brain process. The 
component of NN consists of the neurons, weighted 
connection between neurons, processing unit, and the output 
[18]. Figure 3 shows illustration of NN. X1, X2, X3 is input 
with weight that is denoted by w1, w2, w3, respectively. The 
processing unit to produces output Y consists of two 
functions; sum function (2) and activation function (3). The 
formula of sum function and activation function as follow 
[19]:   
      y = ∑ wjxj + b                                 (2) 
      f(y) = 1 / (1+e-y)         (3) 
 
The explanation of the symbols is 
• xj is an input and wj is weight of input xj 
• y is an output with b as a bias 
 
 
 
 
 
 
Fig. 3. Illustration of neural network 
                                 
  METHODS 
 
This research uses Fertility Data Set from UCI Machine 
Learning Repository with ten attributes of data and a hundred 
sample of semen. The semen sample data have been analyzed 
based on criteria of the WHO 2010. Concentration of the 
sperm are related to life habits, health status, environmental 
factors, and socio-demographic data [15]. Since this data set 
has various attributes, it is considered to be used in this 
research. Table I explains about information of the data set 
and Table II explains about detail of attributes. 
TABLE I.  INFORMATION OF FERTILITY DATA SET 
Item Description 
Area Life 
Number of instances 100 
Number of attributes 10 
Attribute characteristics Real 
Date donated January 17th, 2013 
Characteristics Multivariate 
Associated tasks Classification, Regression 
TABLE II.  DETAIL OF ATTRIBUTES 
Description of 
Attributes Attributes 
Values  
(min – max) Normalized 
Season in which 
the analysis was 
performed 
X1 
(1) winter; (2) 
spring;  
(3) summer; (4) fall 
-1; -0.33; 
0.33; 1 
Age at the time of 
analysis X2 18 – 36 no 
Childish diseases 
(ie. Chicken pox, 
measles, mumps, 
X3 (1) yes; (2) no 0; 1 
978-1-5386-2930-7/17/$31.00 ©2017 IEEE 15-17 November 2017, Melia Purosani Hotel, Yogyakarta, Indonesia 
2017 International Conference on Information Management and Technology (ICIMTech) 
Page 114 
 
 
Description of 
Attributes Attributes 
Values  
(min – max) Normalized 
polio) 
Accident or 
serious trauma X4 (1) yes; (2) no 0; 1 
Surgical 
intervention X5 (1) yes; (2) no 0; 1 
High fevers in the 
last year X6 
(1) less than three 
months ago; (2) 
more than three 
months ago; (3) no 
-1; 0; 1 
Frequency of 
alcohol 
consumption 
X7 
(1) several times a 
day; 
(2) every day; 
(3) several times a 
week; 
(4) once a week; 
(5) hardly ever of 
never 
0.2; 0.4; 0.6; 
0.8; 1 
Smoking habit X8 
(1) never; (2) 
occasional; 
(3) daily 
-1; 0 ; 1 
Number of hours 
spent sitting per 
day 
X9 1 – 16 no 
Diagnosis  Y (N) normal; (O) altered N; O 
  
Furthermore, separable of data is checked using MATLAB. 
The separable of data is needed to know nature of data, linear 
or non-linear. Since the data consist of nine dimensions, then 
based on principle component analysis, it is determined in two 
dimensions based on criteria as follow: 
• First dimension is data with the largest variance. In this 
data, ‘age at the time of analysis’ (X2) has the largest 
variance, since the value of this dimension is nineteen.  
The minimum value of this dimension is eighteen and the 
maximum value is thirty-six. Whenever, others dimension 
consists of two (X3, X4, X5), three (X6, X8), four (X1), 
five (X7), and sixteen (X9) variance. 
• Second dimension is data that is not related with the first 
dimension. Thus ‘season in which the analysis was 
performed’ (X1) is chosen. The reason, there is no relation 
between age and season of analysis. 
Moreover, data is classified by linear classifier and quadratic 
classifier of discriminant analysis using MATLAB. The 
training function is fitcdiscr and the test function is predict of 
MATLAB. 
 Beside discriminant analysis, neural network also used in 
this research as a comparison using MATLAB. The data is 
divided to two groups; training group and testing group 
consists of 67 data and 33 data, respectively. The neural 
network uses 1 hidden layer with three kinds of number of 
neurons; 5 neurons, 10 neurons, and 50 neurons to obtain 
comparison.  
III. RESULT AND DISCUSSION 
The experiment is started from checking of separable of 
fertility data set using MATLAB. The checking process 
involves two attributes; ‘season in which the analysis was 
performed’ (X1) and ‘age at the time of analysis’ (X2). Figure 
3 shows correlation of X1 and X2. Based on mapping result of 
attribute X1 and X2 in figure 3, can be concluded the data is 
non-linear, since the attributes are not separable. Meanwhile, 
Table III shows several fertility data set per attributes to 
illustrate the row data. 
 
 
Fig. 4. Correlation between X1 and X2 
 
TABLE III.  FERTILITY DATA SET 
Data 
# X1 X2 X3 X4 X5 X6 X7 X8 X9 Y 
1 -0.33 0.69 0 1 1 0 0.8 0 0.88 N 
2 -0.33 0.94 1 0 1 0 0.8 1 0.31 O 
3 -0.33 0.5 1 0 0 0 1 -1 0.5 N 
4 -0.33 0.75 0 1 1 0 1 -1 0.38 N 
5 -0.33 0.67 1 1 0 0 0.8 -1 0.5 O 
6 -0.33 0.67 1 0 1 0 0.8 0 0.5 N 
7 -0.33 0.67 0 0 0 -1 0.8 -1 0.44 N 
8 -0.33 1 1 1 1 0 0.6 -1 0.38 N 
9 1 0.64 0 0 1 0 0.8 -1 0.25 N 
10 1 0.61 1 0 0 0 1 -1 0.25 N 
     : 
98 -1 0.67 1 1 1 0 1 -1 0.31 N 
99 -1 0.64 1 0 1 0 1 0 0.19 N 
100 -1 0.69 0 1 1 0 0.6 -1 0.19 N 
TABLE IV.  PERFORMANCE OF DISCRIMINANT ANALYSIS 
Type of discriminant 
analysis Performance Error 
Linear 87% 1 normal and 12 altered 
Quadratic 98% 1 normal and 1 altered 
 
The performance of classification from discriminant 
analysis can be seen in Table IV. Table IV shows comparison 
978-1-5386-2930-7/17/$31.00 ©2017 IEEE 15-17 November 2017, Melia Purosani Hotel, Yogyakarta, Indonesia 
2017 International Conference on Information Management and Technology (ICIMTech) 
Page 115 
 
 
of classification result using linear classifier and quadratic 
classifier from discriminant analysis. Performance of linear 
classifier is 87% with 13 errors of Y attribute consists of 1 
normal and 12 altered, whereas performance of quadratic 
classifier is 98% with 2 errors of Y attribute; 1 normal and 1 
altered. Quadratic classifier has better performance rather than 
linear classifier, not only for performance of classification, 
however in error aspect which quadratic performance has less 
error rather than linear performance. The quadratic classifier 
shows the good performance, since the quadratic estimates 
one covariance matrix per each class. It is more specific, 
whereas the linear estimates one covariance matrix for all 
classes.  
The experiment in neural network using 1 hidden layer 
and consists of three kinds of number of neurons; 5 neurons, 
10 neurons, and 50 neurons. The performance result of neural 
network is shown in Table V. The best performance of neural 
network experiment is experiment with 10 neurons. The 
highest performance of classification is 94% with number of 
errors is the lowest, only 2 errors. However, this performance 
is not better rather than quadratic classifier from discriminant 
analysis. The performance of quadratic classifier is 4% higher 
than the 10 neurons of neural network. This condition can be 
affected by number of neurons or number of hidden layer.  
TABLE V.  PERFORMANCE OF NEURAL NETWORK 
Number of neurons Performance Error 
5 85% 3 normal and 2 altered 
10 94% 2 altered 
50 88% 2 normal and 2 altered 
 
IV. CONCLUSION 
In summary, Quadratic Classifier from discriminant 
analysis shows the best performance in classification of 
multiple attributes data. This performance is influenced by 
classification process from quadratic classifier which 
estimates one covariance matrix per each class. The result of 
fertility data set classification is achieved 98% with the lowest 
errors rather than linear classifier from discriminant analysis 
and neural network. The appropriate method of classification 
produces a good performance of classification. In future work, 
can be considered to use cross validation to validate 
experiment results. 
 
REFERENCES 
 
[1] J. Han, M. Kamber, and J. Pei, Data Mining Concepts and Techniques, 
3rd ed. Elsevier, 2012. 
[2] N. Cristianini and J. Shawe-Taylor, An introduction to support vector 
machines and other kernel-based learning methods. Cambridge 
University Press, 2000. 
[3] R. O. Duda, P. E. Hart, and D. G. Stork, Pattern Classification, 2nd ed. 
Wiley-Interscience, 2012. 
[4] S. R. Safavian and D. Landgrebe, “A Survey of Decision Tree 
Classifier Methodology,” vol. 21, no. 3, 1991. 
[5] D. D. Lewis, “Naive (Bayes) at forty: The independence assumption in 
information retrieval,” Mach. Learn. ECML-98, no. x, pp. 4–15, 1998. 
[6] J. Chen, H. Huang, S. Tian, and Y. Qu, “Feature selection for text 
classification with Naïve Bayes,” Expert Syst. Appl., vol. 36, no. 3, pp. 
5432–5435, 2009. 
[7] J. Ren, S. D. Lee, X. Chen, B. Kao, R. Cheng, and D. Cheung, “Naive 
Bayes Classification of Uncertain Data,” no. 60703110, 2009. 
[8] P. Cunningham and S. J. Delany, “k-Nearest Neighbour Classifiers,” 
Multple Classif. Syst., vol. 34, pp. 1–17, 2007. 
[9] T. Joachims, “Text Categorization with Support Vector Machines : 
Learning with Many Relevant Features,” Mach. Learn. ECML-98, pp. 
137–142, 1998. 
[10] A. Tharwat, “Linear vs . quadratic discriminant analysis classifier : a 
tutorial Alaa Tharwat,” Int. J. Appl. Pattern Recognit., vol. 3, no. 2, pp. 
145--180, 2016. 
[11] T. R. Patil, “Performance Analysis of Naive Bayes and J48 
Classification Algorithm for Data Classification,” Int. J. Comput. Sci. 
Appl. ISSN 0974-1011, vol. 6, no. 2, pp. 256–261, 2013. 
[12] S. D. Jadhav and H. P. Channe, “Comparative Study of K-NN, Naive 
Bayes and Decision Tree Classification Techniques,” Int. J. Sci. Res. 
ISSN (Online Index Copernicus Value Impact Factor, vol. 14611, no. 1, 
pp. 2319–7064, 2013. 
[13] Kalavathi and N. Safar, “Performance Comparison between Naive 
Bayes, Decision Tree and k-Nearest Neighbor,” vol. 9359, no. 6, pp. 
152–161, 2015. 
[14] B. N. Narayanan, R. C. Hardie, and T. M. Kebede, “Analysis of 
Various Classification Techniques for Computer Aided Detection 
System of Pulmonary Nodules in CT,” in Aerospace and Electronics 
Conference (NAECON) and Ohio Innovation Summit (OIS), 2016 IEEE 
National, 2016, pp. 88–93. 
[15] J. L. Girela, J. De Juan, and M. Johnsson, “Predicting seminal quality 
with artificial intelligence methods,” no. November, 2012. 
[16] R. A. Fisher, “The use of multiple measurements in taxonomic 
problems,” Ann. Hum. Genet., vol. 7.2, pp. 179–188, 1936. 
[17] “Mathworks Documentation.” [Online]. Available: 
http://www.mathworks.com/help/stats/discriminant-analysis.html. 
[18] D. Kriesel, A Brief Introduction to Neural Networks. 2007. 
[19] B. Krose and P. Van Der Smagt, An Introduction to Neural Networks, 
no. November. 1996. 
 
 
 


Full paper "Quadratic classifier from discriminant analysis for classification of multiple attributes data: (Case study: Fertility data set)"

Soewito, Benfano

Gaol, Ford L.

Simanjuntak, Echo

Gunawan, Fergyanto E.

Form Reviewer "Smart mobile attendance system using voice recognition and fingerprint on smartphone "

Iskandar, Karto

Warnars, Harco Leslie Hendric Spits

Kosala, Raymondus

Form Reviewer "Software size measurement of knowledge management portal with use case point"

Astriani, Maria Seraphina

Pradono, Satrio

Moniaga, Jurike

Check 18. Delivering an
Interactive Presentation in
supporting if Dynamic Teaching
method with an IT Blueprint.pdf
by Arta Sundjaja
Submission date: 29-Apr-2019 02:57PM (UTC+0700)
Submission ID: 1102509668
File name: n_supporting_if _Dynamic_Teaching_method_with_an_IT_Blueprint.pdf  (232.69K)
Word count: 3738
Character count: 20959






30%
SIMILARITY INDEX
24%
INTERNET SOURCES
13%
PUBLICATIONS
1%
STUDENT PAPERS
1 16%
2 3%
3 3%
4 2%
5 1%
Check 18. Delivering an Interactive Presentation in supporting
if Dynamic Teaching method with an IT Blueprint.pdf
ORIGINALITY REPORT
PRIMARY SOURCES
www.europment.org
Internet  Source
Jurike V. Moniaga, Maria Seraphina Astriani,
Sharon Hambali, Yangky Wijaya, Yohanes
Chandra. "Chapter 56 The Implementation of
MDA Framework in a Game-Based Learning in
Security Studies", Springer Nature, 2018
Publicat ion
borneostudies.org
Internet  Source
Astriani, Maria Seraphina, Satrio Pradono, and
Hoga Saragih. "IT Blueprint for Education
Institution", 2010 Second International
Conference on Advances in Computing Control
and Telecommunication Technologies, 2010.
Publicat ion
spotidoc.com
Internet  Source
media.neliti.com
Internet  Source
6 1%
7 1%
8 1%
9 1%
10 1%
Exclude quotes On
Exclude bibliography On
Exclude matches < 1%
Internet  Source
Hetty Rohayani, Ford Lumban Gaol, Benfano
Soewito, Harco Leslie Hendric. "Estimated
measurement quality software on structural
model academic system with Function Point
Analysis", 2017 International Conference on
Applied Computer and Communication
Technologies (ComCom), 2017
Publicat ion
www.wseas.us
Internet  Source
www.naun.org
Internet  Source
Henny Hendarti, Lianawati Christian, Angelina
Permatasari, Maryani. "The user friendly
features and menu of user satisfaction of
teaching and learning at university in Jakarta",
2016 International Conference on Information
Management and Technology (ICIMTech), 2016
Publicat ion


Check Similarity "Delivering an interactive presentation in supporting of dynamic teaching method with an IT Blueprint framework: IT Initiative-ITBluTric"

Check 35. Software quality
model for Internet of Things
governance.pdf
by Arta Sundjaja
Submission date: 17-May-2019 03:04PM (UTC+0700)
Submission ID: 1102517865
File name: 35._Sof tware_quality_model_f or_Internet_of _Things_governance.pdf  (489.56K)
Word count: 4668
Character count: 26115






8%
SIMILARITY INDEX
7%
INTERNET SOURCES
10%
PUBLICATIONS
7%
STUDENT PAPERS
1 2%
2 2%
3 1%
4 1%
5 1%
Check 35. Software quality model for Internet of Things
governance.pdf
ORIGINALITY REPORT
PRIMARY SOURCES
docplayer.net
Internet  Source
Hetty Rohayani, Ford Lumban Gaol, Benfano
Soewito, Harco Leslie Hendric. "Estimated
measurement quality software on structural
model academic system with Function Point
Analysis", 2017 International Conference on
Applied Computer and Communication
Technologies (ComCom), 2017
Publicat ion
www.emeraldinsight.com
Internet  Source
www.cse.unsw.edu.au
Internet  Source
Wang, Feng, Liang Hu, Jiejun Hu, Jin Zhou,
and Kuo Zhao. "Recent Advances in the
Internet of Things: Multiple Perspectives", IETE
Technical Review, 2016.
Publicat ion
6 1%
7 1%
Exclude quotes On
Exclude bibliography On
Exclude matches < 1%
www.sciencesphere.org
Internet  Source
www.managementjournal.info
Internet  Source


Check Similarity "Software Quality Model for Internet of Things Governance"

Check 34. Software complexity
measurement of water.pdf
by Arta Sundjaja
Submission date: 17-May-2019 03:03PM (UTC+0700)
Submission ID: 1102517599
File name: 34._Sof tware_complexity_measurement_of _water.pdf  (749.6K)
Word count: 3325
Character count: 17489





21%
SIMILARITY INDEX
17%
INTERNET SOURCES
17%
PUBLICATIONS
16%
STUDENT PAPERS
1 5%
2 3%
3 3%
4 2%
5 2%
6 1%
Check 34. Software complexity measurement of water.pdf
ORIGINALITY REPORT
PRIMARY SOURCES
zombiedoc.com
Internet  Source
Hetty Rohayani, Ford Lumban Gaol, Benfano
Soewito, Harco Leslie Hendric. "Estimated
measurement quality software on structural
model academic system with Function Point
Analysis", 2017 International Conference on
Applied Computer and Communication
Technologies (ComCom), 2017
Publicat ion
ijcsi.org
Internet  Source
www.computer-science.nottingham.ac.uk
Internet  Source
onlinelibrary.wiley.com
Internet  Source
Noprianto, Benfano Soewito, Ford Lumban
Gaol, Harco Leslie Hendric, Spits Warnars.
"Simple spreadsheet test case application to
7 1%
8 1%
9 1%
10 1%
11 1%
12 1%
Exclude quotes On
Exclude bibliography On
Exclude matches < 1%
test spreadsheet formula in end-user software
engineering", 2016 11th International
Conference on Knowledge, Information and
Creativity Support Systems (KICSS), 2016
Publicat ion
Submitted to Universiti Malaysia Pahang
Student  Paper
Submitted to Carnegie Mellon University
Student  Paper
Submitted to Curtin University of Technology
Student  Paper
Keyes, . "Function Point Counting Guide",
Software Engineering Handbook, 2002.
Publicat ion
"Table of contents", 2017 International
Conference on Applied Computer and
Communication Technologies (ComCom), 2017
Publicat ion
www.cipl.net.in
Internet  Source
FINAL GRADE
/0
Check 34. Software complexity measurement of water.pdf
GRADEMARK REPORT
GENERAL COMMENTS
Instructor
PAGE 1
PAGE 2
PAGE 3
PAGE 4
PAGE 5


Check Similarity "Software complexity measurement of water poverty mapping application with function point method"

Binus University

Not a member yet